Technical SEO
Building Nested JSON-LD @graph Schema That Connects Your Brand, Founder, and Services
, 3 minute read
A nested JSON-LD @graph is a single script block containing every entity on a page, each with a stable @id, linked to the others by reference. The organization, local business, founder, services and the web page point to one another, so search engines read one coherent description of the brand instead of disconnected fragments.
The problem with schema by plugin
Most sites acquire structured data one plugin at a time. The SEO plugin emits an Organization. The theme adds a LocalBusiness with a slightly different name. A review widget injects its own AggregateRating. Each block validates. None of them knows the others exist.
To a parser these are separate things. It has to infer that the organization on the home page, the local business in the footer and the publisher of the blog are the same entity. Sometimes it infers correctly. You would rather it did not have to.
Give every entity one identity
The fix is the @id. It is a URL that names a thing, by convention the site's address plus a fragment:
https://example.com/#organizationhttps://example.com/#localbusinesshttps://example.com/#founderhttps://example.com/#website
These never change and they appear on every page. A node that needs to mention the organization does not repeat its name and logo. It points to it:
{ "publisher": { "@id": "https://example.com/#organization" } }
One definition, many references.
The four permanent nodes
{
"@context": "https://schema.org",
"@graph": [
{
"@type": "Organization",
"@id": "https://example.com/#organization",
"name": "Example Co",
"founder": { "@id": "https://example.com/#founder" }
},
{
"@type": "LocalBusiness",
"@id": "https://example.com/#localbusiness",
"telephone": "+1-747-555-0100",
"priceRange": "$$$",
"parentOrganization": { "@id": "https://example.com/#organization" }
},
{
"@type": "Person",
"@id": "https://example.com/#founder",
"worksFor": { "@id": "https://example.com/#organization" },
"knowsAbout": ["Technical SEO", "Local SEO"]
},
{
"@type": "WebSite",
"@id": "https://example.com/#website",
"publisher": { "@id": "https://example.com/#organization" }
}
]
}
The Person node matters more than it looks. knowsAbout and worksFor are how you state, in a form a machine can read, who is behind the content and what they are qualified to write about. That is the structural half of what we cover in our piece on E-E-A-T.
Page-level nodes
On top of those, each page adds nodes that describe itself:
- A
WebPagewithisPartOfpointing at the website,aboutpointing at the main subject, andhasPartlisting each anchored section - A
Serviceon service pages, withproviderpointing at the local business - An
Articleon posts, withauthorpointing at the founder andpublisherat the organization - An
ImageObjectfor each meaningful image - A
BreadcrumbList - A
FAQPagewhere the page genuinely contains questions and answers
Because every reference resolves to a node in the same graph, the page describes one connected set of facts. You can see a real one by opening the Inspect Live Schema & Semantics panel on this site from a desktop browser. It parses the graph on whatever page you are viewing.
Rules that keep a graph honest
Mark up only what is on the page. Google's structured data guidelines are explicit: content in markup that users cannot see is a violation. The markup and the semantic HTML5 have to describe the same thing.
Keep facts identical everywhere. The phone number in the graph, in the footer and on the Google Business Profile must be the same string of digits. This is the on-site half of entity co-occurrence.
Do not chase rich results that no longer exist. FAQ rich results are now limited to a small set of government and health sites. The markup still helps systems understand the page. It will not produce the expandable listing it once did.
Generate the graph from data. One source for the business facts, one template per page type. Hand-edited JSON on forty pages drifts within a month.
Validate twice. The Schema Markup Validator checks vocabulary. Google's Rich Results Test checks eligibility. They answer different questions.
Scaling to many locations
A multi-location business repeats the pattern: one Organization, then a LocalBusiness per location, each with its own @id, address, phone, hours and parentOrganization. The location page carries that location's node.
This is what makes it possible for fifty branches to rank as fifty local businesses under one brand, and it is the core of our location-level entity work. Local pages for a single business use the same idea with areaServed, as described under Los Angeles and 747 coverage.
The complete method, including how it ties to page structure, is on our JSON-LD architecture page.
Questions this article raises
Should I use one @graph or several script blocks?
One graph per page is cleaner, because references resolve inside it. Several blocks also work if their @id values match, but they are harder to keep consistent.
Does the @id have to be a real URL that loads?
No. It is an identifier. By convention it is the page or site URL plus a fragment, and what matters is that it is stable and used consistently.
Will this get me a knowledge panel?
Nothing guarantees one. A consistent graph, combined with matching facts on your profile and on other sites, removes the ambiguity that prevents one.
Talk to the person who will actually do the work
No account coordinator, no discovery deck. Call and describe what the phone should be ringing for, or send the site and we'll trace where the signal is leaking.